The Blacklist

In the previous post, I talked about NSFW controls for the Horde and how it's important to allow NSFW creations, but give the tools to individual workers to handle what they are comfortable with.

Now a lot of workers would choose to allow NSFW, but the internet being the internet, they might not want to allow every kind of NSFW creation.

And this is where the blacklists comes in. Three new variables now exist inside your bridgeData.py (so make sure you copy them from the bridgeData_template.py)

blacklist: When this list variable is filled, during job pop, the prompt's words will be compared to this blacklist words, and skipped if any of them matches.

This is a very blunt tool for an elegant job, but it should help take away anything a worker's owner is absolutely sure they don't want to deal with.

censorlist: A bit finer control. Like the blacklist, the censorlist will accept the job and will then compare against the words of the prompt, but instead of rejecting the request outright, it will complete it, but will apply the NSFW censor filter when relevant. This would allow innocent uses of the words inside to go through, while NSFW subjects are hidden.

censor_nsfw: This is similar to the worker's 'censor_nsfw' in the payload. But instead of the worker ensuring they will not have accidental NSFW, this time the worker protects themselves. This variable is only useful when the horde_nsfw variable is set to False. But just because the worker accepts only SFW requests, it doesn't mean that a SFW will not be gory or naughty. This variable ansures that requests which were not appropriately flagged as NSFW, will return a simply black image.

Hopefully with this bit of extra control, horde worker owner can achieve the exact level of control they need to feel comfortable.